I'm surprised Philip Greenspun is acting so naive here. "If it works, why will it take $90 million to keep it running?" This from a computer scientist? Software

I'm surprised Philip Greenspun is acting so naive here. "If it works, why will it take $90 million to keep it running?" This from a computer scientist? Software is never "done", and maintenance on existing computer systems is quite often far more expensive than the initial development. Do you want it to keep updated with correct information about the various plans that are available? To keep up with new tweaks to the regulations? To address new bugs that come up when it's annual renewal time? To stay secure over time? All of these things require a lot of work. And how about fixing the bugs that still remain? Or improving the system? Yes, government projects of this sort are inevitably going to cost far more than a startup company, and mistakes will be made. But $90 million out of the federal budget isn't even worth blinking at. The Defense Department spends that much in about an hour and a half. But if you really want to save money in this area, you should argue for eliminating the exchanges and healthcare.gov altogether. Just sign everyone up for Medicare. Save a few hundred billion over the next few years give or take.